Hi,
Same setup as a few questions earlier: VPN across a ADSL wan
with FRS set on about 40 directories containing 15 GB merely static data.
The target server was low on diskspace but not to the point of error messages (1GB left of 18GB). FRSstaging directory on both sides have plenty of space; content never exceeded 250MB.
Now last week FRS was set on a new directory to replicate, but about 60% of the data in that directory got lost, even subdirectories dissapeared.
And no, it is not what you would initially think: "the user must have switched (empty) target and source, thus overwriting the source with empty data"
As a matter of fact the source was already present on the target. As end result data was deleted from both source and target and even NtFrs_PreExisting___See_EventLog was empty.
That FRS is prone to error under difficult circumstances like low on diskspace is imaginable, but should must be as rubust as just failing to replicate (i.e. just stop), NOT that it deletes data!
imagine you replicate data on say six servers and one has a unforseeable problem, then you could end up losing data on all of them. Beware!
CU
Loek Gijben
Same setup as a few questions earlier: VPN across a ADSL wan
with FRS set on about 40 directories containing 15 GB merely static data.
The target server was low on diskspace but not to the point of error messages (1GB left of 18GB). FRSstaging directory on both sides have plenty of space; content never exceeded 250MB.
Now last week FRS was set on a new directory to replicate, but about 60% of the data in that directory got lost, even subdirectories dissapeared.
And no, it is not what you would initially think: "the user must have switched (empty) target and source, thus overwriting the source with empty data"
As a matter of fact the source was already present on the target. As end result data was deleted from both source and target and even NtFrs_PreExisting___See_EventLog was empty.
That FRS is prone to error under difficult circumstances like low on diskspace is imaginable, but should must be as rubust as just failing to replicate (i.e. just stop), NOT that it deletes data!
imagine you replicate data on say six servers and one has a unforseeable problem, then you could end up losing data on all of them. Beware!
CU
Loek Gijben
